The butt ends are not specially scraped, though they show a tuft of fibre from the takirikiri process of tearing the strips off the butt end of the leaf. The strips are allowed to dry a little, so as to avoid subsequent shrinkage. The whole bundle of strips is divided into two equal parts. The strips composing one part are knotted together at their thin ends, corresponding to the narrower tip end of the leaf, with an overhand knot. The strips are then separated into three equal parts, and, commencing from the knot, are plaited in three-ply braid. As the strips increase in width the band naturally thickens and widens. The plaiting is continued for about 11 in., when the three equally divided parts of the other half of the strips are added, one to each ply. The strips are reversed, the wider butt ends of the added strips being plaited in first. The added strips materially increase the thickness of the band, until the butt ends of the first set are reached, when the band gradually tapers off until the tip ends of the second end are reached and finished off with an overhand knot. The total length of the band I am describing is, 46 in. Its width in the middle is 2 ½ in. and thickness 1 ¼ in. The part of the band between the tapering 11 in. at either end maintained the fairly even width of 2 ½ in., and therefore the business part of the band is roughly 2 ft. in length. It was thus necessary to have two or three bands to encircle the oven, the tapering ends, being too low, being overlapped by the wider-parts of the neighbouring bands. (See Plate 36, fig. 3.)

These braided paepae lasted a long time, and were hung up in the cooking-houses after use. They make a strong serviceable band, but owing to their narrower width the food is more likely to flow over than with the wider-plaited bands. On the other hand, they save the trouble of seeking out fresh flax before cooking each meal.